Urban space-time evolution analysis method based on multi-source historical data

Abstract

The invention discloses an urban space-time evolution analysis method based on multi-source historical data. The method comprises the following steps: 1) acquiring and digitally processing historical city maps, point-of-interest data and realistic paintings in multiple periods; 2) performing syntactic analysis on the urban road network and the water system by adopting a mixed model of topology and angle distance, and calculating core indexes such as integration degree and angle selection degree; 3) based on an optimization algorithm, carrying out normalization processing on the spatial syntactic index and the POI data; 4) extracting dynamic city information from the historical drawing; 5) performing statistical test on dynamic data extracted by drawing and a spatial syntactic analysis result; and 6) analyzing the space-time aggregation mode and function evolution of the public service facility by adopting a location quotient formula and kernel density estimation.By constructing a closed-loop research framework of multi-source data fusion, algorithm optimization and quantitative verification, a highly reliable scientific method is provided for deeply revealing an internal driving mechanism of urban form evolution.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of urban planning, geographic information science, computer vision and history, and particularly relates to a city space-time evolution analysis method based on multi-source historical data. BACKGROUND

[0002] Exploring the evolution of urban form and social characteristics is a core issue for understanding the internal logic of urban development and guiding future urban planning. Traditional research methods rely heavily on literature interpretation, qualitative description of images, or analysis based on a single data source. These methods have significant limitations in comprehensively, objectively, and quantitatively revealing the complex process of urban development. Historical maps and paintings, as valuable visual archives of urban history, contain rich information on spatial structure and human activities. However, their non-structured and non-standardized nature poses significant challenges for quantitative analysis. At the same time, modern urban research requires increasingly precise data, spatial and temporal scales, and multi-dimensional analysis. Traditional methods are inefficient in handling large-scale, multi-period, and multi-source heterogeneous data, and cannot guarantee the scientific objectivity of the conclusions.

[0003] Specifically, the existing technology has the following bottlenecks: 1. Limitations of spatial analysis methods: Space Syntax, as a mature urban network analysis tool, has great potential in assessing urban spatial vitality. However, its classic model still has room for improvement in handling historical maps that carry the cognitive habits of ancient people, especially in fully considering the perception of angle changes by pedestrians (rather than purely topological connectivity).

[0004] 2. Statistical bias of POI data: In the spatial correlation analysis of point of interest (POI) data and street networks, if discrete point data is directly associated with line segments, it is easy to cause high skewness in data distribution, i.e., a small number of streets carry a large number of POIs, while most streets have few POIs. This skewed distribution seriously violates the basic assumption of normality of data in classic statistical methods such as Pearson correlation analysis, which may lead to distorted analysis results and fail to accurately reflect the true relationship between facility distribution and street centrality.

[0005] 3. The absence of cross-source data verification: Although advanced image recognition technologies such as YOLO have been applied in the field of art detection, there is currently a lack of research that systematically uses these technologies to quantitatively extract information on urban social activities from historical paintings and rigorously cross-verify it with spatial structure indicators analyzed from historical maps. This makes the conclusions drawn from a single historical map lack corroboration from another independent information source, raising doubts about their reliability and authenticity.

[0006] 4. The ambiguity of evolution mechanisms: Existing research often fails to construct an integrated analysis framework to fully reveal the persistent influence of urban infrastructure such as water systems and road systems on the development of urban functional areas, as well as the specific mechanisms by which urban social and economic activities such as commercial agglomeration shape historical urban form and social spatial patterns. SUMMARY

[0007] To systematically overcome the above-mentioned shortcomings of existing technologies in studying the spatio-temporal evolution of urban form and social characteristics, the present invention proposes a method for analyzing the spatio-temporal evolution of cities based on multi-source historical data. The core purpose of the present invention is to construct a comprehensive research framework that integrates data fusion, algorithm optimization, quantitative analysis, and cross-verification. This framework aims to scientifically and rigorously integrate and quantitatively analyze urban maps, point of interest (POI) data, and realistic paintings from different historical periods, and to deeply explore and quantify the structural role of water systems and road systems in urban development, as well as how various social and economic activities in cities trigger and respond to changes in urban form and social characteristics.

[0008] The present invention provides the following technical solutions: The method for analyzing the spatio-temporal evolution of cities based on multi-source historical data comprises the following steps: Step S1, data collection and preprocessing: Obtain and digitize at least two historical periods of urban maps, corresponding point of interest (POI) data of public service facilities, and realistic paintings reflecting the urban landscape of the time within the study area, with a span of no less than 50 years for each historical period; Step S2, spatial syntax analysis: Analyze the road and water system network in the digitized urban map using spatial syntax to calculate the integration and angle selection indicators; Step S3, quantification and standardization of public service facilities: Standardize the angle selection indicators and POI data to eliminate data bias and meet the requirements of statistical analysis; Step S4, painting data extraction and verification: Use image recognition technology to extract dynamic urban information from the realistic paintings and spatialize it; Step S5, correlation analysis: the dynamic urban information extracted in step S4 is correlated with the spatial syntax integration and angle selection indexes calculated in steps S2 and S3 to cross-verify the reliability of historical information; Step S6, spatiotemporal evolution analysis of POI data in multiple periods by using location quotient LQ analysis and kernel density estimation KDE; Step S7, comprehensive analysis: the spatial syntax integration and angle selection index characteristics, the results of location quotient LQ analysis and kernel density estimation KDE analysis are integrated to reveal the spatiotemporal evolution mechanism of urban form and social characteristics.

[0009] Further, the standardization processing in step S3 includes: a) logarithmic standardization of the angle selection index ACH to obtain the angle selection standardized index NACH, and the calculation formula is: ; wherein, ACH is the angle selection degree, r is the radius at which the total angle depth is calculated; b) breadth-first algorithm is used for POI data, and weighted summation calculation is performed by combining distance attenuation weight to quantify the number of facilities within the specified topological radius, and then logarithmic transformation is performed to obtain the POI count standardized index NACO.

[0010] Further, the image recognition technology in step S4 is the YOLO deep learning target detection algorithm.

[0011] Further, the correlation analysis in step S5 is Pearson correlation analysis, and the calculation formula is: ; wherein, r is the correlation coefficient, and x and y are sample values of two variables, and are sample means.

[0012] Further, in step S6, the calculation formula of location quotient LQ analysis is: ; wherein, POI is the number of POIs of the region industry, POI is the total number of POIs of all industries in the region , POI is the number of POIs of the entire research area industry, POI is the number of POIs of the entire research area Total POI number for all industries in the entire study area.

[0013] An electronic device includes a memory, a processor, and a graphics processing unit (GPU); the memory is used to store computer executable instructions, and the processor and GPU are used to execute the computer executable instructions.

[0014] A computer readable storage medium having stored thereon computer executable instructions.

[0015] By adopting the above-mentioned technology, compared with the prior art, the beneficial effects of the present application are as follows: 1) The present application realizes scientific fusion and deep quantification of multi-source heterogeneous historical data: the present application innovatively integrates three data sources with different properties, namely historical maps, multi-period POI data and realistic paintings with high artistic value. Through accurate digitization, advanced YOLO image recognition technology and optimized spatial syntax model, objective quantitative extraction and analysis of unstructured historical data (especially paintings) which are difficult to handle traditionally are realized. This greatly enriches the data dimension of urban research and makes up for the lack of depth and breadth caused by single data source or reliance on qualitative analysis in traditional research; 2) An optimized spatial syntax algorithm for the characteristics of historical data is proposed: in view of the characteristics of historical cognitive maps and the inherent complexity of POI data in spatial distribution, the present application designs and verifies an optimized spatial syntax algorithm model. By introducing angle selection normalization (NACH) and POI count normalization (NACO) algorithm, the serious data tilt problem caused by the association of POI data and streets is scientifically solved, so that the processed data is closer to normal distribution, thereby meeting the prerequisite for subsequent reliable statistical analysis. This significantly improves the accuracy, applicability and scientificity of spatial syntax analysis in historical city research; 3) An independent historical information source based cross-validation mechanism is constructed: the present application first systematically applies YOLO algorithm to historical paintings, quantitatively extracts key information (commercial activities and crowd distribution) reflecting urban dynamic vitality from the paintings, and performs Pearson correlation analysis on the spatial syntax analysis results based on historical maps. This step constitutes a cross-validation closed loop, i.e. using an independent historical information source (painting) to verify the conclusions derived from another information source (map), thereby significantly enhancing the reliability, authenticity and scientific persuasiveness of the research results; 4) Provide in-depth study on the mechanism of urban long-term evolution: With the help of various mature spatial econometric models such as location quotient (LQ) analysis and kernel density estimation (KDE), the application can quantitatively evaluate the structural contribution of water system and road system to urban development in different historical periods, and identify the spatial distribution rule, agglomeration mode of different types of public service facilities and the coupling relationship with street accessibility. By comparing and analyzing multi-period data, the historical change trajectory of urban functional center can be clearly outlined, so as to more comprehensively and deeply reveal the internal driving mechanism of the spatio-temporal evolution of urban form and social characteristics; 5) Provide scientific decision support for modern urban planning and heritage protection: The research takes a broader historical and spatial perspective, supplemented by a complete set of quantitative research methods, and systematically discusses the distribution logic and development characteristics of various types of public service facilities in different periods. The research results not only deepen the understanding of the historical evolution of a specific city, but also provide a solid scientific basis for contemporary urban renewal, historical district protection and cultural heritage activation and other planning practices, which is helpful to guide the future urban development to realize sustainability on the basis of respecting historical context. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 is a workflow diagram of a city spatio-temporal evolution analysis method based on multi-source historical data provided by the application; Figure 2 is a visualization diagram of the integration index obtained after spatial syntax analysis of the historical map in the embodiment of the application; Figure 3 is a visualization diagram of the temple type kernel density and integration superimposed heat map obtained after spatial syntax analysis of the historical map in the embodiment of the application; Figure 4 is a visualization diagram of the school type kernel density and integration superimposed heat map obtained after spatial syntax analysis of the historical map in the embodiment of the application; Figure 5 is a visualization diagram of the government type kernel density and integration superimposed heat map obtained after spatial syntax analysis of the historical map in the embodiment of the application; Figure 6 is a schematic diagram of automatically identifying and labeling commercial activities and crowds from historical paintings by using YOLOv5 algorithm in the embodiment of the application. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical scheme and advantages of the application more clear, the application will be further described in detail below in combination with the drawings and embodiments of the specification. It should be understood that the specific embodiments described herein are only used to explain the application and do not limit the application.

[0018] REFERENCE Figure 1The application provides a specific step of a city space-time evolution analysis method based on multi-source historical data as follows: Step 1: Collection and preprocessing of multi-source heterogeneous data Taking Suzhou ancient city as an example, the city map, corresponding point of interest (POI) data of public service facilities and realistic paintings reflecting the city style of three key historical periods (1745, the Qianlong period of the Qing Dynasty, 1938 and 2019) are obtained and digitized. As shown in Table 1: .

[0019] Step 2: Space syntax analysis In order to quantify the spatial structure attributes of Suzhou ancient city, the space syntax theory is adopted, and the Depthmap software is used to analyze the road network and water system network of the three periods. The core calculation indexes are integration and angle selection degree. The integration measures the convenience of each street in Suzhou ancient city to reach all other streets, reflecting its potential as a “destination”; the angle selection degree measures the frequency of the street appearing on the shortest path between any two points, reflecting its “crossing” traffic potential. In order to adapt to the historical cognitive map of Suzhou and more truly simulate the path selection of pedestrians, a hybrid calculation model of topological distance and angle distance is adopted in the research, and all period analyses are carried out in a unified evaluation framework to ensure the scientificity of longitudinal comparison.

[0020] Step 3: Optimization of public service facility quantification and standardization This step is a key technical innovation of the research, aiming to scientifically solve the statistical deviation problem that may occur in the analysis of Suzhou data.

[0021] a) Angle selection standardization (NACH): in order to make the original angle selection degree index (Angular Choice, ACH) calculated from the road network of Suzhou ancient city more consistent with the data distribution requirements of statistical analysis, the following formula (1) is adopted for standardization: Angle selection standardization (NACH): in order to make the original angle selection degree index (Angular Choice, ACH) more consistent with the data distribution requirements of statistical analysis, the application adopts the following formula (1) for standardization: ; In this formula, represents the original angle selection degree within the radius , and represents the total angle depth within the radius . The original The values often present a long-tail distribution, i.e. the range of values is extremely large and highly right-skewed. Directly using such data would violate the assumptions of many statistical models. The log transformation adopted ( ) is a standard variance-stabilizing transformation that can effectively compress the data range and significantly alleviate the skewness of the data, making its distribution closer to normal distribution. Logarithmizing both the numerator and the denominator is a kind of relative standardization, so that the angular selectivity values of different regions or different time periods are comparable.

[0022] b) POI count normalization (NACO): To solve the data skew problem caused by directly associating POIs in Suzhou to street lines, this study designed a breadth-first search (BFS) algorithm using Python programming (with the help of Geopandas and Pandas libraries). This algorithm starts from each street segment in the Suzhou ancient city and searches for reachable POIs within a specified topological radius (e.g. 5 steps). To more realistically reflect the distance decay effect, a weighted summation approach is used, i.e. the farther the POI, the lower the weight (e.g. each step weight decays by 0.9 times). The final weighted count value is then added 1 and logarithmized. This method not only avoids data skew, but also has a more geographically plausible model assumption (distance decay) that conforms to the first law of geography.

[0023] Step 4: Quantitative extraction and cross-validation of painting data This study uses image recognition technology to train a dataset sample of "shop" and "pedestrian" target detection models. This dataset is derived from slicing historical realistic paintings such as "Gusu Prosperous Map" and "Qianlong's Southern Tour Map", and then using LabelImg annotation software to manually frame and label the "shop" (Shop) and "pedestrian" (People) two key elements in the image. These accurately labeled samples form the basis for training and verifying the YOLO-v5 model, aiming to enable the model to automatically recognize and locate these two types of targets in ancient paintings with high precision. From the realistic paintings reflecting the urban landscape of Suzhou in the Qing Dynasty, dynamic urban information such as commercial activities and crowd distribution is extracted (Figure 2), and it is accurately spatialized to the corresponding historical map.

[0024] Step 5: Correlation analysis The core of this step is to cross-validate multi-source data to verify whether the laws revealed by different sources of data are consistent. This invention uses the classic Pearson Correlation Analysis. This method is used to measure the strength and direction of the linear relationship between two continuous variables. Its calculation formula is shown in (2): ; In this formula, and The sample values of two variables (NACH and NACO of a street in Suzhou), and are the sample means of the two variables, is the sample size (the total number of streets). The value ranges from -1 to +1, +1 indicates a complete positive correlation, -1 indicates a complete negative correlation, and 0 indicates no linear correlation. The present application scientifically and quantitatively judges the correlation strength between urban spatial structure and social and economic vitality by testing the correlation between NACH and NACO, NACH and crowd density in paintings, and setting a statistical significance level (usually ). A high and significant correlation coefficient will strongly support the effectiveness of the method proposed by the present application.

[0025] Step 6: Location Quotient (LQ) analysis In order to reveal the evolution of urban functional zoning in different periods, the present application uses the classic index of economic geography, Location Quotient (LQ). LQ is an index used to measure the concentration of a certain industry (here, a certain type of POI) in a specific area relative to a larger reference area (the entire city). Its calculation formula is shown in (3): ; In the application of the present application, refers to the number of POIs of a certain type in a certain area, is the total number of POIs of all types in the area; is the total number of POIs of a certain type in the entire city (reference area N), is the total number of POIs in the entire city. When , it indicates that the concentration level of type facilities in area is higher than the average level of the city, which is the characteristic or dominant function of the area. When , it indicates that the facility is relatively sparse in this area. By calculating and comparing the LQ values in different periods, the rise, decline and spatial migration process of various functional areas in the city can be accurately and quantitatively described.

[0026] Step 7: Kernel Density Estimation (KDE) ​​​The invention adopts Kernel Density Estimation (KDE) to intuitively understand the spatial distribution pattern of various types of public service facilities in the ancient city of Suzhou. KDE can generate smooth and continuous density surfaces, clearly revealing the "hot spot" areas of various facilities in different periods. By superimposing the KDE results with the integration map of spatial syntax, we can further explore the layout preferences of facilities: do they tend to choose the "center" area with the highest accessibility, or do they have other specific location orientations? By superimposing the spatial distribution heat map (KDE) of public service facilities with the structural accessibility (integration R20) of urban road network system, we can intuitively reveal the functional layout logic of Suzhou ancient city in 1745. The overall spatial agglomeration pattern of all types of public service facilities is shown. The figure clearly identifies two main "hot spot" areas in the city, namely the core agglomeration area of public service functions. By classifying and displaying different types of public service facilities (such as temples, education, government agencies, etc.), we can reveal the unique location selection preferences and spatial distribution patterns of various facilities. This atlas not only verifies the geographical accuracy of the calculation model of this study, but also clearly shows that the highly concentrated area of public service facilities at that time is highly consistent with the integration core area analyzed by spatial syntax, confirming the strong guiding effect of urban form structure (high accessibility) on social and economic functions (facility layout) Figure 6 .

[0027] Step 8: Comprehensive analysis and conclusion Through the analysis of historical and modern city map data of Suzhou ancient city by the invention, a representative historical city time and space data set is constructed using YOLO algorithm Figures 2-5 . Based on the aforementioned optimized spatial syntax algorithm, and using Python libraries such as Geopandas and Numpy, the influence of waterway system on the historical development and urban form of Suzhou is analyzed in depth.

[0028] Through the application of multiple spatial econometric models to the map data of Suzhou ancient city in different periods, the following findings are obtained: 1) By controlling the water system as a variable, it is quantitatively confirmed that the water system is an important component in the historical city system of Suzhou, and it plays an important role in urban development and activities, but it has been gradually ignored in modern times. 2) Different types of public service facilities show different accessibility and integration patterns to the water system and road system in different periods. 3) The changes in the distribution of public service facilities correspond to the stage transfer of the historical city center of Suzhou, which is consistent with the phenomenon observed from realistic paintings and historical maps, proving the authenticity and effectiveness of the research method.

[0029] The application discusses the distribution and development characteristics of various public service facilities in different periods of Suzhou ancient city by a wider historical space-time perspective, with the aid of quantitative research method, and reveals the influence of water system and road system on city development, and the historical city form and social change caused by city activities.

[0030] The above merely describes preferred embodiments of the present application and is not used to limit the present application, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for analyzing urban spatio-temporal evolution based on multi-source historical data, characterized in that, The method comprises the following steps: Step S1, data collection and preprocessing: obtaining and digitizing urban maps in at least two historical periods in the study area, corresponding point of interest (POI) data of public service facilities, and realistic paintings reflecting the urban landscape at that time, with a span of not less than 50 years for one historical period; Step S2, spatial syntax analysis: analyzing the road and water network in the digitized urban map using spatial syntax, and calculating the integration degree and angle selection degree index; Step S3, quantification and standardization of public service facilities: standardizing the angle selection degree index and POI data to eliminate data tilt and meet the requirements of statistical analysis; Step S4, painting data extraction and verification: extracting dynamic urban information from the realistic paintings using image recognition technology, and spatializing the dynamic urban information; Step S5, correlation analysis: performing correlation analysis on the dynamic urban information extracted in step S4 and the spatial syntax integration degree and angle selection degree index calculated in steps S2 and S3 to cross-verify the reliability of historical information; Step S6, spatiotemporal evolution analysis of POI data in multiple periods using location quotient (LQ) analysis and kernel density estimation (KDE); Step S7, comprehensive analysis: integrating the spatial syntax integration degree and angle selection degree index characteristics, LQ analysis, and KDE analysis results to reveal the spatiotemporal evolution mechanism of urban form and social characteristics. 2.The method of claim 1, wherein, The standardization processing in step S3 comprises: a) logarithmic standardization of the angle selection degree index ACH to obtain the angle selection standardized index NACH, with the calculation formula being: ； wherein is the angular selectivity, is the radius is the total angular depth at the point b) using the breadth-first algorithm for POI data and combining distance attenuation weight for weighted summation calculation to quantify the number of facilities within a specified topological radius, and then performing logarithmic transformation to obtain the POI count standardized index NACO. 3.The method of claim 1, wherein, The image recognition technology in step S4 is the YOLO deep learning target detection algorithm. 4.The method of claim 1, wherein, The correlation analysis in step S5 is Pearson correlation analysis, with the calculation formula being: ； wherein, is a correlation coefficient, and are sample values of two variables, and are sample means.

5. The urban spatiotemporal evolution analysis method based on multi-source historical and geographical data according to claim 1 is characterized in that: In step S6, the calculation formula of LQ analysis is: ； in, for area The number of POIs in the industry, for The total number of POIs for all industries in the region, For the entire study area The number of POIs in the industry, is the total number of POIs in all industries in the entire study area.

6. An electronic device, comprising: It comprises: a memory, a processor, and a graphics processing unit (GPU); the memory is used to store computer executable instructions, and the processor and GPU are used to execute the computer executable instructions to implement the method of any one of claims 1-5.

7. A computer-readable storage medium, characterized in that, A computer readable storage medium having stored thereon computer executable instructions that, when executed by a processor, can implement the method of any one of claims 1-5.

Images